cassandra-cql 1.0.4 → 1.1.0

data/.travis.yml

@@ -0,0 +1,14 @@
+language: ruby
+services:
+  - cassandra
+rvm:
+  - "1.8.7"
+  - "1.9.2"
+  - "1.9.3"
+#  - jruby-18mode # JRuby in 1.8 mode
+#  - jruby-19mode # JRuby in 1.9 mode
+  - ree
+#  - rbx-18mode
+#  - rbx-19mode
+# uncomment this line if your project needs to run something other than `rake`:
+# script: bundle exec rspec spec

data/README.md

@@ -0,0 +1,81 @@
+# Cassandra [![Build Status](https://secure.travis-ci.org/kreynolds/cassandra-cql.png)](http://travis-ci.org/kreynolds/cassandra-cql)
+The Apache Cassandra Project (http://cassandra.apache.org) develops a highly scalable second-generation distributed database, bringing together Dynamo's fully distributed design and Bigtable's ColumnFamily-based data model.
+
+# CQL
+Cassandra originally went with a Thrift RPC-based API as a way to provide a common denominator that more idiomatic clients could build upon independently.
+However, this worked poorly in practice: raw Thrift is too low-level to use productively, and keeping pace with new API methods to support (for example) indexes in 0.7 or distributed counters in 0.8 is too much for many maintainers.
+
+CQL, the Cassandra Query Language, addresses this by pushing all implementation details to the server; all the client has to know for any operation is how to interpret "resultset" objects.
+So adding a feature like counters just requires teaching the CQL parser to understand "column + N" notation; no client-side changes are necessary.
+
+(CQL Specification: http://cassandra.apache.org/doc/cql/CQL.html)
+
+# Quick Start
+
+## Establishing a connection
+
+    # Defaults to the system keyspace
+    db = CassandraCQL::Database.new('127.0.0.1:9160')
+
+    # Specifying a keyspace
+    db = CassandraCQL::Database.new('127.0.0.1:9160', {:keyspace => 'keyspace1'})
+
+    # Specifying more than one seed node
+    db = CassandraCQL::Database.new(['127.0.0.1:9160','127.0.0.2:9160'])
+  
+## Creating a Keyspace
+
+    # Creating a simple keyspace with replication factor 1
+    db.execute("CREATE KEYSPACE keyspace1 WITH strategy_class='org.apache.cassandra.locator.SimpleStrategy' AND strategy_options:replication_factor=1")
+    db.execute("USE keyspace1")
+
+## Creating a Column Family
+
+    # Creating a column family with a single validated column
+    db.execute("CREATE COLUMNFAMILY users (id varchar PRIMARY KEY, email varchar)")
+
+    # Create an index on the name
+    db.execute("CREATE INDEX users_email_idx ON users (email)")
+
+## Inserting into a Column Family
+
+    # Insert without bound variables
+    db.execute("INSERT INTO users (id, email) VALUES ('kreynolds', 'kelley@insidesystems.net')")
+
+    # Insert with bound variables
+    db.execute("INSERT INTO users (id, email) VALUES (?, ?)", 'kway', 'kevin@insidesystems.net')
+  
+## Updating a Column Family
+
+    # Update
+    db.execute("UPDATE users SET email=? WHERE id=?", 'kreynolds@insidesystems.net', 'kreynolds')
+  
+## Selecting from a Column Family
+
+    # Select all
+    db.execute("SELECT * FROM users").fetch { |row| puts row.to_hash.inspect }
+      {"id"=>"kway", "email"=>"kevin@insidesystems.net"}
+      {"id"=>"kreynolds", "email"=>"kreynolds@insidesystems.net"}
+
+    # Select just one user by id
+    db.execute("SELECT * FROM users WHERE id=?", 'kreynolds').fetch { |row| puts row.to_hash.inspect }
+      {"id"=>"kreynolds", "email"=>"kreynolds@insidesystems.net"}
+
+    # Select just one user by indexed column
+    db.execute("SELECT * FROM users WHERE email=?", 'kreynolds@insidesystems.net').fetch { |row| puts row.to_hash.inspect }
+      {"id"=>"kreynolds", "email"=>"kreynolds@insidesystems.net"}
+  
+## Deleting from a Column Family
+
+    # Delete the swarthy bastard Kevin
+    db.execute("DELETE FROM users WHERE id=?", 'kway')
+
+# Notes
+
+## Changing Validation on Columns with existing/unvalidatable data
+
+  If you have existing data and change the validation on a column in an incompatible
+  way (ie. blank strings with a column validated as Integer), a CastException will be raised.
+  The exception has a 'bytes' attribute that will give you access to the bytes that caused the problem.
+  
+  Other columns in a row can still be accessible via index or column_name without raising that exception.

data/cassandra-cql.gemspec

@@ -6,7 +6,7 @@ Gem::Specification.new do |s|
   s.version     = CassandraCQL::VERSION
   s.platform    = Gem::Platform::RUBY
   s.authors     = ["Kelley Reynolds"]
-  s.email       = ["kelley@insidesystems.net"]
+  s.email       = ["kelley.reynolds@rubyscale.com"]
   s.homepage    = "http://code.google.com/a/apache-extras.org/p/cassandra-ruby/"
   s.summary     = "CQL Interface to Cassandra"
   s.description = "CQL Interface to Cassandra"
@@ -15,7 +15,12 @@ Gem::Specification.new do |s|
   s.rubyforge_project         = "cassandra-cql"
 
   s.add_development_dependency "bundler", ">= 1.0.0"
-  s.add_development_dependency "rcov", ">= 0.9.9"
+
+  if RUBY_VERSION >= "1.9"
+    s.add_development_dependency "simplecov"
+  else
+    s.add_development_dependency "rcov", ">= 0.9.9"
+  end
   s.add_development_dependency "rspec", ">= 2.6.0"
   s.add_development_dependency "rake", ">= 0.9.2"
   s.add_development_dependency "yard", ">= 0.7.2"

data/lib/cassandra-cql.rb

@@ -16,7 +16,7 @@ limitations under the License.
 
 module CassandraCQL; end;
 unless CassandraCQL.respond_to?(:CASSANDRA_VERSION)
-  require "cassandra-cql/1.0"
+  require "cassandra-cql/1.1"
 end
 
 here = File.expand_path(File.dirname(__FILE__))

data/lib/cassandra-cql/1.1.rb

@@ -0,0 +1,7 @@
+module CassandraCQL
+  def self.CASSANDRA_VERSION
+    "1.1"
+  end
+end
+
+require "#{File.expand_path(File.dirname(__FILE__))}/../cassandra-cql"

data/lib/cassandra-cql/1.1/result.rb

@@ -0,0 +1,6 @@
+module CassandraCQL
+  module V11
+    class Result < CassandraCQL::Result
+    end
+  end
+end

data/lib/cassandra-cql/1.1/statement.rb

@@ -0,0 +1,6 @@
+module CassandraCQL
+  module V11
+    class Statement < CassandraCQL::Statement
+    end
+  end
+end

data/lib/cassandra-cql/database.rb

@@ -32,6 +32,7 @@ module CassandraCQL
       }.merge(thrift_client_options)
 
       @keyspace = @options[:keyspace]
+      @cql_version = @options[:cql_version]
       @servers = servers
       connect!
       execute("USE #{@keyspace}")
@@ -41,6 +42,7 @@ module CassandraCQL
       @connection = ThriftClient.new(CassandraCQL::Thrift::Client, @servers, @thrift_client_options)
       obj = self
       @connection.add_callback(:post_connect) do
+        @connection.set_cql_version(@cql_version) if @cql_version
         execute("USE #{@keyspace}")
         @connection.login(@auth_request) if @auth_request
       end

data/lib/cassandra-cql/result.rb

@@ -104,10 +104,12 @@ module CassandraCQL
           end
         end
       else
-        if (row = fetch_row).kind_of?(Fixnum)
-          {row => row}
-        else
-          row.to_hash
+        if row = fetch_row
+          if row.kind_of?(Fixnum)
+            {row => row}
+          else
+            row.to_hash
+          end
         end
       end
     end
@@ -130,4 +132,4 @@ module CassandraCQL
       end
     end
   end
-end
+end

data/lib/cassandra-cql/row.rb

@@ -20,12 +20,27 @@ module CassandraCQL
     
     def initialize(row, schema)
       @row, @schema = row, schema
+      @value_cache = Hash.new { |h, key|
+        # If it's a number and not one of our columns, assume it's an index
+        if key.kind_of?(Fixnum) and !column_names.include?(key)
+          column_name = column_names[key]
+          column_index = key
+        else
+          column_name = key
+          column_index = column_names.index(key)
+        end
+        
+        if column_index.nil?
+          # Cache negative hits
+          h[column_name] = nil
+        else
+          h[column_name] = ColumnFamily.cast(@row.columns[column_index].value, @schema.values[@row.columns[column_index].name])
+        end
+      }
     end
   
     def [](obj)
-      column_index = obj.kind_of?(Fixnum) ? obj : column_names.index(obj)
-      return nil if column_index.nil?
-      column_values[column_index]
+      @value_cache[obj]
     end
 
     def column_names
@@ -35,7 +50,9 @@ module CassandraCQL
     end
   
     def column_values
-      @values ||= @row.columns.map { |column| ColumnFamily.cast(column.value, @schema.values[column.name]) }
+      @row.columns.map { |column|
+        @value_cache[column.name]
+      }
     end
   
     def columns

data/lib/cassandra-cql/statement.rb

@@ -74,7 +74,7 @@ module CassandraCQL
         obj.map { |member| quote(member) }.join(",")
       elsif obj.kind_of?(String)
         "'" + obj + "'"
-      elsif obj.kind_of?(Fixnum) or obj.kind_of?(Float)
+      elsif obj.kind_of?(Numeric)
         obj
       else
         raise Error::UnescapableObject, "Unable to escape object of class #{obj.class}"

data/lib/cassandra-cql/types/abstract_type.rb

@@ -15,6 +15,17 @@ limitations under the License.
 =end
 
 module CassandraCQL
+  module Error
+    class CastException < Exception
+      attr_reader :bytes
+
+      def initialize(message = nil, bytes = nil)
+        super(message)
+        @bytes = bytes
+      end
+    end
+  end
+
   module Types
     class AbstractType
       def self.cast(value)
@@ -31,6 +42,8 @@ module CassandraCQL
           int = int - (1 << bytes.length * 8)
         end
         int
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to int", bytes)
       end
 
       def self.bytes_to_long(bytes)
@@ -41,6 +54,8 @@ module CassandraCQL
         else
           val
         end
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to long", bytes)
       end
     end
   end

data/lib/cassandra-cql/types/ascii_type.rb

@@ -18,7 +18,9 @@ module CassandraCQL
   module Types
     class AsciiType < AbstractType
       def self.cast(value)
-        RUBY_VERSION >= "1.9" ? value.to_s.dup.force_encoding('ASCII-8BIT') : value.to_s.dup
+        RUBY_VERSION >= "1.9" ? value.to_s.dup.force_encoding('US-ASCII') : value.to_s.dup
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to ascii", value)
       end
     end
   end

data/lib/cassandra-cql/types/boolean_type.rb

@@ -18,7 +18,12 @@ module CassandraCQL
   module Types
     class BooleanType < AbstractType
       def self.cast(value)
+        # Do not assume that an empty string is false
+        raise if value.empty?
+
         value.unpack('C') == [1]
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to boolean", value)
       end
     end
   end

data/lib/cassandra-cql/types/date_type.rb

@@ -19,6 +19,8 @@ module CassandraCQL
     class DateType < AbstractType
       def self.cast(value)
         Time.at(bytes_to_long(value) / 1000.0)
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to Date", value)
       end
     end
   end

data/lib/cassandra-cql/types/decimal_type.rb

@@ -19,6 +19,8 @@ module CassandraCQL
     class DecimalType < AbstractType
       def self.cast(value)
         BigDecimal.new(bytes_to_int(value[4..-1]).to_s) * BigDecimal.new('10')**(bytes_to_int(value[0..3])*-1)
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to decimal", value)
       end
     end
   end

data/lib/cassandra-cql/types/double_type.rb

@@ -18,7 +18,12 @@ module CassandraCQL
   module Types
     class DoubleType < AbstractType
       def self.cast(value)
+        # Do not return nil, this should be an error
+        raise if value.empty?
+
         value.unpack('G')[0]
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to double", value)
       end
     end
   end

data/lib/cassandra-cql/types/float_type.rb

@@ -18,7 +18,12 @@ module CassandraCQL
   module Types
     class FloatType < AbstractType
       def self.cast(value)
+        # Do not return nil, this should be an error
+        raise if value.empty?
+
         value.unpack('g')[0]
+      rescue
+        raise Error::CastException.new("Unable to convert bytes to float", value)
       end
     end
   end

data/lib/cassandra-cql/types/utf8_type.rb

@@ -19,6 +19,8 @@ module CassandraCQL
     class UTF8Type < AbstractType
       def self.cast(value)
         RUBY_VERSION >= "1.9" ? value.to_s.dup.force_encoding('UTF-8') : value.to_s.dup
+      rescue => e
+        raise CassandraCQL::Error::CastException, "Unable to convert bytes to UTF8: #{bytes.inspect}", caller
       end
     end
   end

data/lib/cassandra-cql/types/uuid_type.rb

@@ -19,6 +19,8 @@ module CassandraCQL
     class UUIDType < AbstractType
       def self.cast(value)
         UUID.new(value)
+      rescue => e
+        raise CassandraCQL::Error::CastException, "Unable to convert bytes to UUID: #{value.inspect}", caller
       end
     end
     

data/lib/cassandra-cql/version.rb

@@ -15,5 +15,5 @@ limitations under the License.
 =end
 
 module CassandraCQL
-  VERSION = "1.0.4"
+  VERSION = "1.1.0"
 end

data/spec/column_family_spec.rb

@@ -62,7 +62,7 @@ describe "ColumnFamily class" do
 
     it "should turn a packed long into a number" do
       number = 2**33
-      packed = [number >> 32, number].pack("N*")
+      packed = [number >> 32, number % 2**32].pack("N*")
 
       ColumnFamily.cast(packed, "LongType").should eq(number)
       ColumnFamily.cast(packed, "CounterColumnType").should eq(number)
@@ -70,7 +70,7 @@ describe "ColumnFamily class" do
 
     it "should turn a packed negative long into a negative number" do
       number = -2**33
-      packed = [number >> 32, number].pack("N*")
+      packed = [number >> 32, number % 2**32].pack("N*")
 
       ColumnFamily.cast(packed, "LongType").should eq(number)
       ColumnFamily.cast(packed, "CounterColumnType").should eq(number)

data/spec/misc_spec.rb

@@ -0,0 +1,68 @@
+require File.expand_path('spec_helper.rb', File.dirname(__FILE__))
+include CassandraCQL
+
+describe "Miscellaneous tests that handle specific failures/regressions" do
+  before(:each) do
+    @connection = setup_cassandra_connection
+    @connection.execute("DROP COLUMNFAMILY misc_tests") if @connection.schema.column_family_names.include?('misc_tests')
+    @connection.execute("CREATE COLUMNFAMILY misc_tests (id text PRIMARY KEY)")
+  end
+
+  context "with ascii validation" do
+    before(:each) do
+      @connection.execute("ALTER COLUMNFAMILY misc_tests ADD test_column ascii")
+    end
+    
+    it "should be consistent with ascii-encoded text" do
+      @connection.execute("INSERT INTO misc_tests (id, test_column) VALUES (?, ?)", 'test', 'test_column').should be_nil
+      row = @connection.execute("SELECT test_column FROM misc_tests WHERE id=?", 'test').fetch
+      row['test_column'].should eq 'test_column'
+      @connection.execute("INSERT INTO misc_tests (id, test_column) VALUES (?, ?)", 'test', row['test_column']).should be_nil
+      row = @connection.execute("SELECT test_column FROM misc_tests WHERE id=?", 'test').fetch
+      row['test_column'].should eq 'test_column'
+    end
+  end
+
+  context "with unvalidatable data" do
+    before(:each) do
+      @connection.execute("INSERT INTO misc_tests (id, good_column, bad_column) VALUES (?, ?, ?)", 'test', 'blah', '')
+      @connection.execute("ALTER COLUMNFAMILY misc_tests ADD bad_column int")
+      @row = @connection.execute("SELECT good_column, bad_column FROM misc_tests WHERE id=?", 'test').fetch
+    end
+    
+    it "should have valid column_names" do
+      @row.column_names.should eq ['good_column', 'bad_column']
+    end
+    
+    it "should have raise an exception with column_values" do
+      expect { @row.column_values }.to raise_error(Error::CastException)
+    end
+    
+    it "should be able to fetch good columns even with bad columns in the row" do
+      @row['good_column'].should eq 'blah'
+      @row[0].should eq 'blah'
+    end
+
+    it "should only cast a column once regardless of it's access method" do
+      expect {
+        @row['good_column'].should eq 'blah'
+        @row[0].should eq 'blah'
+      }.to change {
+        @row.instance_variable_get(:@value_cache).size        
+      }.by 1
+    end
+
+    it "should throw an error trying to fetch the bad column by name" do
+      expect { @row['bad_column'] }.to raise_error(Error::CastException)
+    end
+
+    it "should throw an error trying to fetch the bad column by index" do
+      expect { @row[1] }.to raise_error(Error::CastException)
+    end
+
+    it "the raw bytes should be accessible for repair" do
+      @row.row.columns[1].name.should eq 'bad_column'
+      @row.row.columns[1].value.should eq ''
+    end
+  end
+end